The truth behind the Yanny/Laurel mystery

The Yanny/Laurel mystery has been solved - kind of.

Earlier this week the world was split into two opposing camps - but instead of fighting over religion, borders or resources, the disagreement was over whether a short audio clip was of a person saying 'Yanny' or 'Laurel'.

But it's emerged neither is correct - it wasn't a person at all, but a speech synthesiser.

The New York Times tracked down the original source of the auditory meme to Roland Szabo, an 18-year-old student from Georgia, who found the clip on website vocabulary.com. He posted it on web forum Reddit, and it spread online from there.

Mr Szabo said he was looking up the world 'Laurel' when a disagreement broke out amongst his classmates as to whether the voice was saying 'Laurel' or 'Yanny' - much like how in 2015, the infamous blue/gold dress sparked debates that continue to this day (it was eventually revealed the dress was in fact blue).

Much of the debate appears to have been fuelled by the fact the version of the 'Laurel' clip that went viral in a tweet by popular social media personality Cloe Feldman has significantly less bass in it than the original found on vocabulary.com.

The brains of people who hear 'Yanny' appear to be locking onto those higher frequencies.

In a short video, Twitter user Dylan Bennett used audio software to remove certain frequencies from the recording. When he took out the high frequencies, the word became much closer 'Laurel', but sounded remarkably like 'Yanny' when the bass frequencies were cut.

"Each sound is made up of several frequencies, and those that create 'Yanny' are higher than those for 'Laurel', cognitive neuroscientist Lars Riecke of Maastricht University told tech site The Verge.

Older people are less likely to hear the higher frequencies, so will likely lean towards 'Laurel', whilst younger listeners may think they're hearing 'Yanny'.

It could also depend on the speakers being used to listen to the recording. In the original comment thread on Reddit, some listeners reported changing their minds after switching to a different computer.

"I thought it must be a joke," wrote MagickMaggie, who initially heard 'yanny'. "How could Yeah sound like Lore? How could annie sound like oral? I switched to my laptop and suddenly, all I could hear was a deep male voice saying, 'laurel.'"

"At first I heard Yanny and then listened for Laurel enough times to hear Laurel," wrote user snowdisc. "Then there was a period where it sounded like someone was saying Yanny and someone else was saying Laurel at the same time. But now I can only hear Laurel."

The New York Times posted a tool which allows the user to slide between high and low frequencies, making it easy to turn 'laurel' into 'Yanny'.

Another factor that could be contributing to the confusion is that the recording isn't of a real human being, so has qualities our ears consider unnatural. As psycholinguist Suzy Styles explained on Twitter, spectrograms of the audio - visual representations of the present frequencies, and how strong they are - show there are several dark patches known as 'formants'.

"In normal speech, there are three dark bands under 5000Hz," she wrote. "Now, in the Yanny/Laurel clip there are more than three of these formant bands. So unless this speaker had two completely separate tongues, this ambiguous speech has been carefully crafted to fool the ears."

Finally, the brain is very good at filling in missing information, so what starts out as an ambiguous sound will be translated into something approaching a real word by our subconscious.

"Our brains 'fill in' missing information, and how that happens seems to vary a lot from one person to the next," psychology professor Andrew Oxenham of the University of Minnesota told Live Science.

The dress debate of 2015 was fuelled by the washed-out colours in the image fooling people's brains into thinking it was a white and gold dress.

Newshub.